Back to the early days of the web with Docker and Jupyter Notebook

Since I love all things virtual, I really looked forward to the video conversation between Stephen Downes, the facilitator of our course E-learning 3.0 (#el30), and Tony Hirst from the Open University in the UK about virtual boxes and containers.

The Open University is all about distance learning. People bring their own devices, only constrained by modest minimum specifications. Using technologies such as containers more things become possible and life for students and professors becomes easier.

Hirst and Downes talked about Docker, used to run software packages called “containers”. As Wikipedia explains, containers are isolated from each other and bundle their own tools, libraries and configuration files; they can communicate with each other through well-defined channels.

The other interesting product they discussed is Jupyter notebook,a web-based interactive computational environment for creating “Jupyter notebooks documents”. Jupyter combined with Docker allows for powerful
digital media assets which contain their own code, which can be modified by the user. One can change any aspect of a program and run it, without the need for installing software.

In the video you’ll see Hirst demonstrating it with music code. This evolution is like revisiting the beginning of the web, when you could easily see the source code of documents and understand it, so one could modify it without too much trouble.

But is this technology something a single user, not necessarily a coder, could set up? We’ll have to find out. Maybe we’ll experiment with Docker and Jupyter during the #el30 course. Personally, I’d love to stand in a virtual environment, VR-enabled, using a virtual screen to interact with digital assets powered by virtual software environments.

Starting a new blog: learning with Moocs

The ongoing course E-learning 3.0 (#el30) inspired me to start another blog, Learning with Moocs. I’ll focus there on a number of technical experiments regarding the decentralized web. Read my first post on the new blog for more information. On MixedRealities I’ll keep posting about the ideas behind connected learning, decentralized web, virtual environments and telepresence as ways to enhance collaboration.

An Experience API for learning everywhere (also in virtual worlds)

I never heard about Experience API (xAPI) but now I did thanks to a video conversation between education expert Stephen Downes and Shelly Blake-Plock, Co-Founder, President and CEO of Yet Analytics in the course #el30. xAPI is a new specification for learning technology that makes it possible to collect data about the wide range of experiences a person has, online and offline.

It captures data in a consistent format about a person or a group’s activities from many different technologies. xAPI uses a simple vocabulary and lots of things can be recognized and communicated with this API: mobile learning, simulations, virtual worlds, serious games, real-world activities, all kinds of learning. Learning activities are being captured in secure statements to learning record store (LRS). The LRS is a server that is responsible for receiving, storing, and providing access to learning records.It can help to develop learning analytics, dashboards, visualizations and recommendations useful for learners.

One of the important questions is whether this technology can be made useful and controlled by the learner or whether it remains in the realm of institutions and companies.

One of the inspiring things Shelly said: “Our experiences and activities become media. Not only when we upload stuff to YouTube but also behind the scenes to make sure internet applications work like we want them to.” It reminds me of the information philosopher Luciano Floridi who describes our lives as being “onlife”.

Media

I’m a journalist, not an educator, but in fact this discussion remains relevant to me. In media we increasingly use realtime audience analytics. A simple news website will track not only how many clicks a story gets, but also how long people read or watch it, how much of a text they read. A/B testing shows which title, picture or caption “works best”.

These metrics can be used to check the learning of journalists and editors: how efficient are they in bringing a story, but at least as interesting would be to get an idea about the learning by the reader or the viewer. Did they actually learn something they can use in their lives, or did they get at least new insights? The readers could integrate metrics of their media consumption into a tool which would point out whether it benefits them in terms of taking better decisions, or in having more interesting conversations with fellow citizens. Media users would own the data and use user-friendly apps to analyze their learning experiences.

Journalism and education seem to be very similar. Learners and audiences become active participants. Monopolies tumble as there are many ways to get learning or information. Business models are put into question. Education and learning explore the possibilities of data, data science, artificial intelligence and have to deal with all-important questions about privacy and autonomy.

Here is the interview with Shelly:

Federated social networks

I’ve been exploring Mastodon which is like a community-owned and ad-free Twitter. It’s an open source social network. There is no one person, company or server running it. People can create their own version of Mastodon, which is called an instance. The person, group or community doing so can determine which are the rules to be followed on that particular instance. Even though the instances are privately owned, users can still communicate with others on other instances.

This means that Mastodon cannot go bankrupt and disappear. An instance can stop because of lack of financing or because the person/community involved give it up, but the other instances would not go down because of that. There are more than a million registered users now.

The design of Mastodon is very user-friendly, Tweetdeck-like. Talking about Twitter, there’s also a way to find Twitter-friends using Mastodon.

Mastodon uses the standard protocol ActivityPub and is part of Fediverse (federation & universe) which basically is a bunch of interoperable federated social networks: Mastodon, diaspora, Friendica, Hubzilla, GNU Social, Socialhome, GangGo, postActiv, Pleroma, Misskey, PeerTube, Osada.

Of course, also in this universe there are moderation problems, fake news messages, and lots of discussions about the best technological routes to follow. Outside of the universe people point out that the commercial social networks are so much bigger. Even so, federated social networks seem to grow rapidly and they are fascinating – every instance being a bit like a virtual world. You can follow me on @rolandlegrand@mastodon.social.

Maybe I’ll create my very own instance, but then I’ll have to decide whether to use a hosted solution or to do the hosting myself, running my own server – something we’ll discuss in more detail in the E-Learning 3.0 course (#el30).

Learning, Connectivism, AI, Virtual Environments

I’m still thinking about the video-conversation of Stephen Downes and George Siemens. Maybe we could change the title of the course from the rather bland ‘E-Learning 3.0, Distributed Learning Technology’ to ‘Human Learning in the Age of Machine Learning’.

After all, Connectivism, the learning theory developed by Stephen en George, is about “the idea that knowledge is essentially the set of connections in a network, and that learning is the process of creating and shaping those networks.”
Wikipedia explains that “learning (defined as actionable knowledge) can reside outside of ourselves (within an organization or a database), is focused on connecting specialized information sets, and the connections that enable us to learn more are more important than our current state of knowing”.Connectivism sees knowledge as a network and learning as a process of pattern recognition.

Which means, as is explained in the video, that in times of rapidly developing Machine Learning Connectivism is a very suitable learning theory. It blends philosophy, educational practices and technological skills. It emphasizes the ability to make decisions and to choose what to learn, connecting with others and thus empathizing with those others. The theory is also related with the Extended Mind ideas of the philosopher Andy Clark.

Virtual Reality and Virtual Environments enhance human connections, they turn telepresence into much more than videoconferencing: they enable people who are dispersed all over the planet to share the same space. Downes mentions Virtual Reality in his slide show “Personal Learning versus Personalized Learning“, but I’d love to elaborate on the possibilities and the many aspects involved in VR and learning. Not only humans benefit from virtual environments, simple virtual worlds are also used to train artificial intelligences to acquire a basic knowledge about the world (like elementary physics) which make those intelligences grow in their world-knowledge. There are many other aspects: for those interested in distributed networks, the virtual world High Fidelity (optimized for VR) uses a distributed architecture, a cryptocurrency and it registers digital assets using a blockchain.

In 2008 we discussed Connectivism with a small group of learners in Second Life. It’s not too late to organize similar experiences for this course.

What and how to learn and teach in times of Artificial Intelligence?

The education theorists, practitioners and technologists George Siemens and Stephen Downes united again for the course E-Learning 3.0. Stephen was in a hotel room in Toronto and George somewhere in Australia, but the wonders of YouTube made them unite (after a search for the light switches).

In my earlier posts I referred to the very first MOOC they facilitated (the very first MOOC in general) in 2008, Connectivism & Connected Knowledge (CCK08). There were about 2,000 participants, these days it seems harder to get that many people. Reasons might be the marketing power of Coursera, edX and similar big platforms, the fact that big social media (Facebook) seemed to make blogging and RSS-feeds less relevant.

What else changed? In the video-discussion george Siemens mentions Artificial Intelligence (AI). If machine learning can learn about everything humans can learn, why would we still learn? One part of the answer is that as humans, we cannot not learn. But what is uniquely human? Is it, as he says, compassion and kindness? The ‘beingness’ of humans?

Stephen is not convinced. ML could develop ethics too. But maybe the way we experience the world as biological organisms is different from the way an AI can be aware. So humans could be the voice in the AI’s mind telling that there are more ways to look at the world.

If humans cannot not learn, maybe we should think about teaching. Learning at school can be a frustrating experience, and maybe what we require students to learn is not suited in the age of AI. Stephen point out that the capacity to take decisions and to choose by the learner will become even more important. That was obvious already in 2008 when the learners got an avalanche of learning materials to digest during CCK08 and they were told a that time already to pick and choose. Other important aspects, which are not being measured by universities, is the ability to contemplate about our place in the universe and in the community.

Also in the video: an interesting conversation about fake news and blockchain. Attention: the real conversation starts after about 8 minutes. 

How much marketing does a MOOC need?

Stephen Downes’ MOOC about distributed learning, E-Learning 3.0, had a pre-start last week – which I missed, but I watched the recording.

It seems Stephen was a bit disappointed – only a few people turned up for the first live-video, maybe this was just because many thought the course would start on October 15.

Or maybe there are more fundamental reasons. When Stephen en George Siemens facilitated a very first MOOC in 2008, blogs and RSS-readers still were important tools. There was no competition from Coursera, edX, Udemy and other huge and heavily financed platforms. Marketing is all-important in online education these days (and in everything else).

Or is it? Maybe people really look for alternatives. Maybe it does not matter whether you attract 35 participants and not 100,000 – as long as the interaction is deep and creative. But then again I’d personally prefer a rich interaction attracting 100,000, and that huge group would split up into many smaller networks working on their own thing but being inspired by the global MOOC.

We’ll find out much more in the coming weeks.

A MOOC about distributed e-learning

I’m learning about the decentralized web these days. I even made a very simple site using The Beaker Browser and files in the dat-format. On that site I keep track of my adventures in decentralized land.

I’m pretty excited to learn that the education expert Stephen Downes is launching a MOOC about distributed e-learning. He’ll talk about IPFS and other technologies as well. Here is an older post of his about these topics (and many other things).

The course starts on October 15 and I’ll be reporting about it on this blog.I’m not employed in education but in journalism, but serious journalism is a form of education for all those concerned. So I consider myself very much a learner and I always enjoyed what Stephen Downes builds, organizes and facilitates. His newsletter is extremely inspiring.

New media theatre

So many interesting things are happening in and around the virtual world High Fidelity that I can’t keep up. The company raised $35 million in June and combines virtual worlds, virtual reality and the blockchain. Philip Rosedale, creator of both Second Life and High Fidelity, also manages to invite interesting people for in-world talks, and in the aftermath of the capital increase he had a fascinating talk with Charlie Fink, an expert in VR, AR, new media and a columnist at Forbes. He also is the author of an AR-enabled book, Metaverse.

In this video Rosedale explains a bit more about the future of High Fidelity and he and Fink brainstorm about new theatre and movie formats which would convert the spectators into actors – a bit like roleplaying in virtual worlds. All this, like concepts such as volumetric video, is rather new to me. During the discussion I heard about other experiments in virtual environments where spectators were converted into bubbles who could follow the actor around. It is obvious there will be formidable challenges like managing the huge data flows involved and finding an equilibrium between the freedom of roleplaying and the need for narrative structure.

Virtual Reality and “Reality”

I was at Les Rencontres de la Photographie in Arles, France. They also have a VR Arles Festival with a competition for VR productions. Winner is this year Treehugger: Wawona by the London design studio Marshmallow Laser Feast. One becomes a water particle and travels into a giant sequoia tree from the roots to the top. The production also won the Storyscapes Award at the Tribeca Film Festival 2017. The installation involves haptic elements (a backpack, gloves, and a fake tree).

The same studio also made In the Eyes of the Animal where you explore the woods through the eyes of different animals. Which made me think of what the VR pioneer Jaron Lanier says about VR, that it makes you see reality in a new way. Philosophers talk about a ‘multiplicity of worlds’ and VR allows us to experience that. Reality is plural.